Correct conversion between URL and filename paths.

doc/changelog.txt

@@ -5,6 +5,9 @@ Fixes:
   Closes: SF bug #3196917
 - checking: Prevent false positives when detecting the MIME type
   of certain archive files.
+- checking: Correct conversion between file URLs and encoded
+  filenames. Fixes false errors when handling files with Unicode
+  encodings.
 - cmdline: Fix filename completion if path starts with ~
 
 Changes:

linkcheck/checker/fileurl.py

@@ -69,7 +69,7 @@ def get_os_filename (path):
     """Return filesystem path for given URL path."""
     if os.name == 'nt':
         path = prepare_urlpath_for_nt(path)
-    return fileutil.pathencode(urllib.url2pathname(path))
+    return urllib.url2pathname(fileutil.pathencode(path))
 
 
 def is_absolute_path (path):
@@ -146,7 +146,7 @@ class FileUrl (urlbase.UrlBase):
         if self.is_directory():
             self.set_result(_("directory"))
         else:
-            url = fileutil.pathencode(self.url)
+            url = urllib.url2pathname(fileutil.pathencode(self.url))
             self.url_connection = urllib2.urlopen(url)
             self.check_case_sensitivity()
 

tests/checker/data/dir.result

@@ -4,8 +4,8 @@ real url file://%(curdir)s/%(datadir)s/dir/
 name %(datadir)s/dir
 valid
 
-url %%ED%%BB%%AD%%AF%%BF.dat
-cache key file://%(curdir)s/%(datadir)s/dir/%%ED%%BB%%AD%%AF%%BF.dat
-real url file://%(curdir)s/%(datadir)s/dir/%%ED%%BB%%AD%%AF%%BF.dat
-name í»­¯¿.dat
+url %%C3%%AD%%C2%%BB%%C2%%AD%%C2%%AF%%C2%%BF.dat
+cache key file://%(curdir)s/%(datadir)s/dir/%%C3%%AD%%C2%%BB%%C2%%AD%%C2%%AF%%C2%%BF.dat
+real url file://%(curdir)s/%(datadir)s/dir/%%C3%%AD%%C2%%BB%%C2%%AD%%C2%%AF%%C2%%BF.dat
+name í»­¯¿.dat
 valid